Pento is a team of Machine Learning and Software Engineers who design, build, and deploy AI systems that solve real business problems. From semantic search and recommendation engines to computer vision and generative AI, we deliver end-to-end — from research and prototyping through production.
Our engineering team brings deep hands-on expertise across ML, full-stack development, and cloud infrastructure. We integrate directly into client teams and operate in North American time zones.
We contribute tools and research back to the community, with a focus on AI-powered search and ML workflows:
- minicoil-en-es — miniCOIL EN-ES, co-built with Qdrant: a sparse retriever that matches English and Spanish in a single index, with no translation step. Model on Hugging Face.
- alphaXiv MCP — With alphaXiv: an MCP server that brings alphaXiv's research corpus into Claude Code, Cursor, and other AI coding tools. See from research paper to production.
- ml-ralph — An autonomous ML research agent that runs experiments, maintains a living plan, and learns as it goes.
- vectory — Track and compare embedding versions. Essential tooling for teams iterating on search and retrieval systems.
- streamlit-terran-timeline — A Streamlit component that builds face-recognition timelines for videos.
